python-notebook-devlisted

# Python Notebook Development ## Purpose Create reproducible `# %%` Python notebooks for focused experiments. Use classes by default for stateful steps, I/O, pipelines, and lifecycle; use pure functions for small stateless transforms. Keep notebooks executable from a clean kernel and ready to migrate into `python-code-dev`. ## Workflows | Command | Action | |---|---| | `/python-notebook-dev new <notebook-name>` | Create the next numbered experiment notebook | | `/python-notebook-dev adapt <notebook.py>` | Bring an existing notebook under these rules | | `/python-notebook-dev migrate <notebook.py>` | Move reusable logic into `python-code-dev` | | `/python-notebook-dev verify <notebook.py>` | Execute and inspect reproducibility/validation | For `new`, run `${AGENTS_SKILLS_DIR}/scripts/new.sh "<notebook-name>"` when available. It creates a notebook under the current project and refuses unsafe overwrites. ## Deterministic Naming New notebooks use `notebooks/NNN-slug/notebook_name.py`: - `NNN` is one greater than the highest existing three-digit folder prefix; start at `001`. - If the candidate exists, increment until free. Never reuse or overwrite an experiment number. - Folder slug: lowercase; spaces/underscores become `-`; remove other characters; collapse hyphens; trim; max 50 chars. - File stem: safe lowercase `snake_case`, max 50 chars. Example: `Customer Churn EDA` → `notebooks/001-customer-churn-eda/customer_churn_eda.py`. ## Environment Use `uv` for every Pytho